Short-term prediction of groundwater level using improved random forest regression with a combination of random features

Xuanhui Wang,Xilai Zheng,Hui Peng,Jia Xin,Tailian Liu,Bo Zhang

doi:10.1007/s13201-018-0742-6

Xuanhui Wang, Xilai Zheng + Show 4 more

Open Access

https://doi.org/10.1007/s13201-018-0742-6

Copy DOI

Abstract

To solve the problem where by the available on-site input data are too scarce to predict the level of groundwater, this paper proposes an algorithm to make this prediction called the canonical correlation forest algorithm with a combination of random features. To assess the effectiveness of the proposed algorithm, groundwater levels and meteorological data for the Daguhe River groundwater source field, in Qingdao, China, were used. First, the results of a comparison among three regressors showed that the proposed algorithm is superior in terms of forecasting variations in groundwater level. Second, the results of experiments were used to show the comparative superiority of the proposed method in terms of training time and complexity of parameter optimization. Third, using the proposed algorithm, the highest prediction accuracy was achieved by employing precipitation P(t − 2), temperature T(t), and groundwater level H(t) as the best time lag. This improved random forest regression model yielded higher accuracy in forecasting the variation in groundwater level. The proposed algorithm can also be applied to cases involving low-dimensional data.

Highlights

Fluctuations in groundwater level (GWL) can be used to evaluate groundwater stability and flow as well as the characteristics of the aquifer
Based on the above issues, in this paper, we propose a method of GWL modeling based on canonical correlation forests with a combination of random features (CCF-CRF) that expands the low-dimensional feature vector space to highdimensional space and uses canonical correlation components for oblique splits
A regression tree (RT), where each non-leaf node contains a set of decision rules and each leaf node is the outcome of a prediction, is a form of decision tree (DT) (Quinlan 1993; Rodriguez-Galiano et al 2014)

Summary

Introduction

Fluctuations in groundwater level (GWL) can be used to evaluate groundwater stability and flow as well as the characteristics of the aquifer. Groundwater is the main source of farmland irrigation in major agricultural regions. Most government agencies collect GWL data once or twice a year in major agricultural areas, this is not sufficient for short-term studies. It is necessary to achieve acceptable prediction accuracy of GWL variations when previous information is not available and the computational sources are limited. It is well known that the dynamic variation in GWL is influenced by meteorological phenomena, urbanization, tidal effects, and land subsidence (Khaki et al 2015). Of these, such meteorological parameters as atmospheric pressure, frost, precipitation,

Methods

Results

Conclusion